Decision tree based on cloud model and its application in slope stability prediction

Slope stability prediction is a common problem in engineering as it always bring disasters to human. This paper aims to solve this problem using the method in machine learning. First, we decide the number of language values by searching the slope training data or experiences. Second, in order to obtain the parameters of each cloud model of language values, boundary value method is proposed. Third, after each cloud model is constructed according to its parameters, slope training data are all converted to language values using these cloud models. Decision tree is constructed based on the slope training data which are described by language values. At last, this decision tree is used for slope stability prediction. In our experiments, 73 slope instances are used to learn, and 9 slope instances are used to test the model. Experimental results prove that this method is very practical and effective. © 2010 Binary Information Press.